OpenAI emailed advertisers about a new ad unit test: several ads shown together inside one ChatGPT prompt response. The trial covers only a small share of eligible placements for now, according to Search Engine Roundtable, which reported the announcement.
The format, which OpenAI describes as “presenting multiple relevant ads together within a single placement,” extends the existing ChatGPT ad unit. Those placements run on a second-price auction, so the winner pays one increment above the next-highest qualifying bid rather than their full maximum. The relevance score (which factors in context signals, creative copy, and landing-page fit) influences which ads qualify and how they rank within the placement, so bid alone does not determine position.
For practitioners tracking AI search monetization, the structural implication is direct. A generative answer surface that previously accepted one ad per prompt is now being tested as multi-unit inventory. This mirrors the trajectory that Google’s AI Mode and AdSense-adjacent placements have been following: the more a query resolves inside the AI layer, the more that layer itself becomes the ad placement, not the blue-link page below it.
ChatGPT Ads supports two buying types: CPM (Reach), with a default maximum of $60 per thousand impressions, and CPC (Clicks), where OpenAI recommends a maximum bid of $3 to $5. The multi-advertiser test applies to both, though advertisers can now set custom CPM max bids, a change from the earlier fixed-bid structure. The ability to customize bids means brands with tighter efficiency targets can participate without committing to the platform’s prior ceiling.
OpenAI also announced several updates to the ChatGPT Ads Manager alongside the placement test:
- Campaign budget type can now be switched from lifetime to daily.
- CPM campaigns can be cloned and converted to CPC with a single click.
- Bulk edits to campaigns, ad groups, and ads are now available directly in the Ads Manager interface (previously CSV-only).
- Daily budgets will operate as average daily targets with weekly pacing flexibility, consistent with how major ad platforms handle budget smoothing.
- Targeting is expanding beyond the United States, Canada, Australia, and New Zealand to include the United Kingdom, Japan, South Korea, Brazil, and Mexico.
The country expansion is notable for search teams managing international campaigns. Demand-side presence in the UK, Japan, and Brazil gives advertisers a reason to build ChatGPT-specific creative and landing-page variations rather than treating the platform as a US-only test.
Brands already working to appear in AI-generated answers should treat the second-price auction structure as a cost-efficiency signal: optimizing for relevance (contextual fit, landing-page alignment, creative clarity) can lower effective cost-per-impression without raising bids. Search teams that have been tracking AI answer citation patterns have a transferable skill here. The same signals that help content surface organically in ChatGPT responses, specificity, clear topical authority, and matched intent, are the signals that lift relevance scores in the ad auction.
The multi-advertiser test is early and limited in scope. The Ads Manager improvements suggest OpenAI is building toward a more mature self-serve platform regardless of how the placement test scales.
Reported by Search Engine Roundtable (Barry Schwartz) in early June 2026.
